Aeolian Grain Transport

Developments in the modelling of turbulent wind over hills and sand dunes of different shapes by Hunt et al. [1], Carruthers et al. [2] are briefly described, and compared with earlier studies of Jackson and Hunt [3] and Walmsley et al. [4]. A new model (FLOWSTAR) is described; it has a more accurate description of airflow close to the surface, which is not in general logarithmic at typical measurement heights. Comparisons are made between the new model and the results of non-linear models using higher-order turbulence schemes, especially for surface shear stress.

The widely predicted and observed drop in velocity and shear stress at the base of a dune is confirmed by FLOWSTAR. It is clear that common models for the saltation flux based only on µ_* are not appropriate at the toe of the dune where they predict a piling-up of sand.

Comparisons of the wind speed are made between the model and different sets of measurements over a dune, by Howard and Walmsley [5], and by our group in recent field measurements over a dune in Oman, and in a new wind-tunnel study of Howard’s dune. It is found that the FLOWSTAR calculations agree well with these sets of measurements upwind of the brink. Since the profile is not logarithmic over the dune at the measurement heights, estimates of µ_* from wind measurements over dunes are likely to be less accurate than the FLOWSTAR computation of µ_*. The saltation flux was measured over the Oman dune and increases in proportion to computed value of µ_* ³ over the dune. This supports the use of the Lettau and Lettau version of Bagnold’s flux formula for modelling sand transport over the most of the upwind slopes of sand dunes.

A mathematical-physical model for the effect of erosion and deposition on the temporal and spatial variation of the size distributions of sediments is proposed and investigated. The model is closely related to the logarithmic hyperbolic distribution and preserves the hyperbolic shape of log-size distributions. In order to facilitate specification of the model in applications the concept of a ‘local erosion time’ is introduced. The local erosion time gives the local time scale of grain sorting changes. Explicit formulae for the temporal-spatial variation of the hyperbolic parameters predicted by the model are derived for stable bed forms and alluvial streams. The results are compared to field observations. Also the effect of the spatial gradient of the grain sorting is considered.

The location-scale invariant parameters χ and ξ of the hyperbolic distribution have a triangular domain of variation which is referred to as the hyperbolic shape triangle. There are close analogies between the hyperbolic distributions, the generalized logistic distributions and the beta distributions, which make it possible also to define a logistic or a beta shape triangle.

Using a population concept of sediment samples it is shown that both the hyperbolic shape triangle and the beta shape triangle provide useful information on variations of grain size distributions, in cross-shore transects. This information can be related to the process-oriented erosion/deposition model developed previously in connection with the hyperbolic shape triangle.

Applying the population concept on more than a thousand beach sand samples from the southern Israeli Mediterranean coast allowed abandonment of the common single sample approach. The study suggests six distinct grain size populations related to a sequence of six beach environments, entitled: 0. and 1. inshore; 2. step; 3. mid-swash zone; 4. top-swash zone; 5. backshore; 6. and D. farbackshore wind blown sand and coastal dunes. The invariant hyperbolic shape parameter Q indicates a gradual change of the grain-size distributional form across the left part of the hyperbolic shape triangle. The shape positions (χ, ξ) in the triangle suggest that the subaqueous populations are subjected mainly to - erosion-deposition processes and move along one of the upper ’hammock’ curves in the shape triangle. The subaerial populations were found to follow mainly ℵ-erosion-deposition processes and to move along the -0.6 to -0.7 Q lines. However, this group is divided into two basically different depositional environments: water-lain backshore populations dominated by the swash-backwash bidirectional sheet flow and wind-blown sand originating from the backshore sediments. The different grain-size populations reflect the sum of modes of transport which dictate the grain size cutoffs and the typical grain size, thus defining the shape of the distributions.

Experimental results are presented which illustrate the effects of permanent surface obstructions on saltation phenomena. It is shown that the topographic drift geometry and the dimensionless erosion rates of windward erosion associated with cylindrical obstacles are strong functions of the cylinder aspect ratio. For short cylinders, there is also significant erosion taking place in the far wake. These two erosional areas develop due to different sets of separation vortex systems. For multi-element roughness arrays, sparse array data are presented which illustrate the increase of threshold friction speed with element frontal area density and roughness element drag coefficient.

The objective of this research is to assess the relationship among measurements of roughness parameters derived from radar backscatter, the wind, and topography on various natural surfaces and to understand the underlying physical causes for the relationship. This relationship will form the basis for developing a predictive equation to derive aerodynamic roughness (z₀) from radar backscatter characteristics. Preliminary studies support the existence of such a relationship at the L-band (24 cm wavelength) direct polarization (HH) radar band frequencies. To increase the confidence in the preliminary correlation and to extend the application of the technique to future studies involving regional aeolian dynamics, the preliminary study has been expanded by: 1) defining the empirical relationship between radar backscatter and aerodynamic roughness of bare rocks and soils, 2) investigating the sensitivity of the relationship to microwave parameters using calibrated multiple wavelength, polarization, and incidence angle aircraft radar data, and 3) applying the results to models to gain an understanding of the physical properties which produce the relationship. The approach combines the measurement, analysis, and interpretation of radar data with field investigations of aeolian processes and topographic roughness.

The nature of interactions between surface winds and natural desert surfaces has important implications for aeolian sediment transport. We report here initial results from measurements of boundary layer wind profiles and surface roughness at 5 sites in Death Valley, U.S.A. and discuss their implications. The sites studied were a flat to gently undulating gravel and sand reg, three alluvial fan surfaces, including one with a well-developed desert pavement, and a silt and clay playa. Aerodynamic roughness estimates range from 0.00018 to 0.00537 m and increase in parallel with the visual estimates of topographic roughness at each site. Microtopography was measured with template and laser profiling devices. The standard deviation of surface elevations (RMS height) appears to provide a good index of surface roughness. It correlates well with field observations of topographic roughness and aerodynamic roughness estimates.

A portable wind erosion tunnel has been used to measure the wind erodibility of nine soil types with a range of surface textures under two treatments (bare uncultivated and bare cultivated) in western New South Wales. The erodibility has been characterised by the soil flux function Q(u _*), whereQis the streamwise soil flux (measured with a modified Bagnold soil trap) and u _* the friction velocity (obtained by fitting the logarithmic wind profile law to a wind profile measured in the tunnel). Threshold friction velocities u _*t were also observed for the range of surfaces. These data represent the only Australian tests published to date and supplement the American measurements of semi-arid soils by Gillette [1], [2], [3].

Averaging over 10 replicate plots for each surface type was necessary to smooth the large scatter in Q and smaller scatter in u _*. The Q values spanned three decades of magnitude. Soils with a sandy loam surface texture were the boundary between the highly erodible sand and the basically noner-odible clay. Cultivation increased the erodibility of the majority of soils by about a factor of 10, but decreased the erodibility of the clay. The function Q(u _* is well described by the Owen [4J soil flux equation. Threshold friction velocity decreased as soil texture became sandier. In comparison with work of Gillette [3], Australian soils have lower u _*t values, which is most likely due to higher sand and lower silt contents.

From South Mauritania to Sudan between the 150 and 600 mm isohyets the sub-Saharan semi-arid lands are covered by a fixed sand sea. Two areas are discussed in this paper: in Niger west and east of the river Niger bend from N’Guigmi to Niamey and in Mali leeward of the Niger’s inland delta. The aim is 1) to describe the morphometric data of the sand accumulation, 2) to understand the actual aeolian removal of particles, 3) to analyse the effects of wind erosion in desertification.

The results obtained show that the deposits south of Sahara accumulated on a surface of 600 km are due to the same Wind Action System (WAS). The grain size is constantly and globally decreasing in the Harmattan wind direction. The wind abrasion on the sand particles is increasing in the same direction. The grain size 400 µm is a threshold: the upper limit of actual, average aeolian deflation.

This area, which previously had a positive sediment balance, during the last drought combined with agricultural overuse became an area of sand and dust removal with a negative sediment balance.

The effect of sea cliffs on aeolian sand encroachment inland was monitored by means of traps measuring the sand transport and microanemometers measuring wind flow, because these are affected by the topography of the cliffs. Measurements were carried out on two cliffs in the southern coastal-plain of Israel: one 160 m and the second 240 m long, with front slope inclination of 30° to 40° and a height of 22-25 m. Results show that sand is incapable of climbing even moderate cliff slopes of 10°-15° because storm wind impinges on the cliffs at angles of 45° to 50° to the rim of the cliffs and is diverted to flow in a helical pattern parallel to the shoreline along the front slope of the cliffs. The diverted sand-moving wind is again deflected at the northern end of the cliffs, where its magnitude abates and it thus deposits sand that is being carried along the beach. The flow of the wind on the beach parallel to the foot of the cliff represents the effect of a non-homogeneous secondary wind. The wind is accelerated along the cliff, and as a result sand transport increases along the cliff. After a length of 150-200 m along the front slopes of the cliffs a state of equilibrium is achieved.

The coastal dunes of Aquitaine number more than 1500, and have a total included volume of between 10 and 20 X 10⁹ m³. There are four major types of dunes: parabolic dunes, barchans, round dunes and the foredune. All the dunes were constructed by westerly winds, blowing off the sea. Dating of these dunes is based on paleosols of the Great Dune of Pilat, the highest dune in Europe. The base of this dune is dated at about 3500 years B.P., but most of the dunes were probably formed during the last 2000 years. These dunes obstructed a number of small rivers inducing the development of lakes lie behind them. One small estuary was transformed into a lagoon, the Arcachon lagoon.

The sands which built the dunes come from the Atlantic coastal beaches. The present rate of coastal retreat is about 1 to 2 m a year. The coastal erosion rates now observed can be used to estimate the rate of intrusion of sand inland from the coast. Measurements with sand traps and by cartographic comparison give an eastward aeolian flux estimated of between 15 and 30 m³ a year per m length of foredune.

The basal part of the Middle Buntsandstein (Lower Triassic) on Helgoland in the southern North Sea is composed of lacustrine redbeds with thin white beds of aeolian sand sheet origin. The aeolian deposits formed during lake low-stands, and sand sheet formation was caused by a restricted sand supply perhaps in combination with frequent wind changes and high-speed wind events. With time aeolian sedimentation was progressively influenced by brief flood events or by minor lake transgressions. Ultimately the sand sheets were covered by fine-grained lake deposits during a new phase of lake high-stand. Preservation of the aeolian deposits records the overall low-energy of the littoral processes, and early cementation. The aeolian deposits define sedimentary cycles of various thickness, and it is suggested that the cyclicity records Early Triassic orbital climatic forcing.

Field observations following a major storm surge on 26 February 1990 on the Sefton Coast in Northwest England showed that the rate of aeolian sand transport and backshore dune construction varied significantly alongshore in response to variations in beach morphology and sand wetness. These variations, in turn, were determined by longshore variations in the character of marine sediment transport processes and the distribution of subsurface silt and peat formations which outcrop locally on the foreshore. The grain size of the beach sands was found to be very uniform and was not a factor influencing spatial variations in sand transport rate. Since the beach sediments consisted of well-sorted and very well-sorted fine sands, partly reworked from the dunes during the storm surge, deflation was not grain-size selective during the period of post-surge dune recovery.